Skip to content

Conversation

@jinhongyii
Copy link
Contributor

LowBatchGemv rule only apply to function with spatial symbolic var, so that this rule skips matmul in attention kernel which has symbolic_var kv_seq_len

@jinhongyii
Copy link
Contributor Author

@tvm-bot rerun

2 similar comments
@jinhongyii
Copy link
Contributor Author

@tvm-bot rerun

@jinhongyii
Copy link
Contributor Author

@tvm-bot rerun

Copy link
Contributor

@MasterJH5574 MasterJH5574 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

@jinhongyii
Copy link
Contributor Author

@tvm-bot rerun

@jinhongyii jinhongyii force-pushed the fix-low-batch-gemv branch from f397ced to 5df1fbe Compare March 8, 2024 07:06
@tqchen tqchen merged commit 5bbe1ab into apache:main Mar 9, 2024
Lunderberg pushed a commit to Lunderberg/tvm that referenced this pull request Mar 12, 2024
thaisacs pushed a commit to thaisacs/tvm that referenced this pull request Apr 3, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants